AI Ensemble Learning for Cardiovascular Risk Stratification Using Retinal Imaging and Non-traditional Factors
Care Process & Redesign
Technology
Singapore Healthcare Management Congress
SingHealth
11 February 2026
To improve accessibility to cardiovascular risk screening by using AI on lifestyle factors and retinal images, without the. AI-enabled cardiovascular screening with retinal imaging is feasible, implementable, and cost-effective.
Year Submitted: 2025
Published Date: 11 February 2026
Tags: Technology, Care Process ＆ Redesign, Automation, Artificial Intelligence, Productivity, Cost Saving, Digitalisation
About this Content
Aims
To improve accessibility to cardiovascular risk screening by using AI on lifestyle factors and retinal images, without the need for blood tests, and to enhance detection accuracy and early treatment initiation.
Background
Traditional methods like the Framingham Risk Score have limitations in accessibility and accuracy. AI-enabled methods using retinal imaging and lifestyle factors offer a promising alternative.
Methods
1563 patients were recruited and underwent standardized questionnaires and retinal photography. An ensemble machine learning method was used, incorporating deep neural networks and other models, with data divided into training and validation sets.
Results
AI models performed better than the Framingham Risk Score, especially in diabetics where only retinal images were needed. The incremental cost-effectiveness ratio was S$9,681 with 104 QALY gained.
Conclusion
AI-enabled cardiovascular screening with retinal imaging is feasible, implementable, and cost-effective.
Lessons Learnt
Different models are needed for diabetics and non-diabetics. Retinal data alone is sufficient for diabetics, while non-diabetics require additional lifestyle and demographic data.
Keywords
AI, cardiovascular, retinal imaging, screening, cost effectiveness
Innovators' Details
Innovators' Details
Healthcare Cluster(s) | SingHealth |
Organization(s) Involved | National Heart Centre Singapore, Singapore National Eye Centre, National University of Singapore, Duke -NUS Health Services and Systems Research |
Platform(s) | Singapore Healthcare Management Congress |
Healthcare Professional Group(s) | Medical |
Applicable Specialty or Discipline | Cardiology |
Project Lead(s) | Weiting HUANG |
Project Member(s) | Xiayan SHEN |
Connect with this contributor!
Weiting HUANG - singaporehealthcaremanagement@singhealth.com.sg